R is a software environment for statistical computing and graphics. Using R you can do rigorous statistical analysis, clean and manipulate data, and create publication-quality graphics.
clustering map
Packages are programs that you import into R to help make tasks easier. The most popular R packages for working with data include dplyr, stringr, tidyr, and ggplot2.
There’s no easy way (yet) for new R users to find R packages that they might need. People are working on this problem. In the meantime, consult the following list or ask a Librarian!
Resources include:
You can create graphs in R without installing a package, but packages will allow you to create better visualizations that are any of the following:
ggplot2 is the most popular visualization package for R. It’s the best all-purpose package for creating many types of 2-dimensional visualizations.
Highcharter is an R package known as an htmlwidget, which allows you to use popular javascript packages for visualization and create interactive web charts. Highcharter is the R interface to the popular highchartsJS, a charting library in javascript. It’s free to use highcharter unless you are using it for a commercial or government purpose.
data(citytemp)
hc <- highchart() %>%
hc_xAxis(categories = citytemp$month) %>%
hc_add_series(name = "Tokyo", data = citytemp$tokyo) %>%
hc_add_series(name = "London", data = citytemp$london) %>%
hc_add_series(name = "Other city",
data = (citytemp$tokyo + citytemp$london)/2)
hc
Leaflet is popular among web developers for creating interactive web maps. It’s an htmlwidget for R based on LeafletJS.
m <- leaflet(options = leafletOptions(zoomControl = FALSE, dragging=FALSE, minZoom = 15, maxZoom = 15)) %>%
addTiles() %>% # Add default OpenStreetMap map tiles
addMarkers(lng=-78.6697, lat=35.7876,
popup="Hello World!")
m # Print the map
Plotly is somewhat new and is making inroads among data scientists for making interactive visualizations.
p <- plot_ly(economics, x = ~date, y = ~unemploy / pop)
p
ggplot2 was created on the principles of the Layered Grammar of Graphics (2010), by Hadley Wickham and based of off work from Wilkinson, Anand, & Grossman (2005) and Jaques Bertin (1983).
Essentially: graphs are like sentences you can construct, and they have a grammar. The grammar of graphics consists of the following:
at least one layer:
plus the following: * scale
* coordinate system
* facet (optional)
These components make up a graph.
Download the following file: script.R Click the blue download button
Open RStudio. File > Open File…
Select the script.R file that you just downloaded (probably in your Downloads folder) Click Open
Let’s see an example of a simple graph created with ggplot. We are going to use the mpg data set about different cars and their properties.
?mpg to learn more about this dataset. To run the code, highlight it and then click Run. (shortcut keys: Mac: command + Enter, Windows: CTRL + Enter)?mpg
head(mpg) to see the first few rows of the data.head(mpg)
## # A tibble: 6 x 11
## manufacturer model displ year cyl trans drv cty hwy fl
## <chr> <chr> <dbl> <int> <int> <chr> <chr> <int> <int> <chr>
## 1 audi a4 1.8 1999 4 auto(l5) f 18 29 p
## 2 audi a4 1.8 1999 4 manual(m5) f 21 29 p
## 3 audi a4 2.0 2008 4 manual(m6) f 20 31 p
## 4 audi a4 2.0 2008 4 auto(av) f 21 30 p
## 5 audi a4 2.8 1999 6 auto(l5) f 16 26 p
## 6 audi a4 2.8 1999 6 manual(m5) f 18 26 p
## # ... with 1 more variables: class <chr>
The graph below uses ggplot2 to look for correlation between a car’s engine displacement and highway mileage.
library(ggplot2): loads the ggplot2 library
ggplot() : function that tells R that you want to make a graph with ggplot
data = mpg : says that you want to use the mpg dataset (sample data that comes with R)
geom_point(): function that says you want to make a scatterplot
mapping = aes(): function that allows you to map data variables to X and Y axes
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy))
Make a scatterplot with cyl mapped to the x-axis and hwy mapped to the y-axis.
ggplot(data= mpg) + geom_point(mapping = aes(x=cyl, y=hwy))
Make a scatterplot of disp=x and hwy=y with class mapped to the color aesthetic. Run:
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy, color = class))
The type of drive system the car has (4-wheel, rear-wheel, and front-wheel) is mapped to color.
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy, color = drv))
Variables can be mapped to the following aesthetic parameters. If you are publishing in b/w, and can’t use color, you might want to use size or shape:
colorsizeshapealpha - transparencySubstitute another aesthetic in place of color. Run the code:
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy, color = drv))
Facets are a way to create multiple smaller charts, or subplots, based on a variable. Run this code to see what faceting does:
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy)) +
facet_wrap(~ class, nrow = 2)
Substitute class for another variable in the dataset. Ex: trans, drive, or cyl
ggplot(data = mpg) +
geom_point(mapping = aes(x = displ, y = hwy)) +
facet_wrap(~ class, nrow = 2)
Facet grids allow for an extra dimension of faceting. Run this code in your script to see what facet_grid() does:
ggplot(data = mpg) + geom_point(mapping = aes(x = displ, y = hwy)) +
facet_grid(class ~ cyl)
Now create a new scatter plot with the dataset diamonds using ggplot2. Refer to previous code examples for assistance.
head(diamonds)
## # A tibble: 6 x 10
## carat cut color clarity depth table price x y z
## <dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
## 1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
## 2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
## 3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
## 4 0.29 Premium I VS2 62.4 58 334 4.20 4.23 2.63
## 5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
## 6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
ggplot(data = diamonds) +
geom_point(mapping = aes(x = carat, y = price, color=cut)) + facet_wrap(~cut, nrow=2)
So far we have just worked with one chart layer. But it’s possible to add more layers to charts in ggplot2, and style those layers individually if you want to. Here’s an example using geom_smooth(), which fits a model to the data. Notice that the color variable is only applied to the scatter points, and not the line.
ggplot(data = diamonds) +
geom_point(mapping = aes(x = carat, y = price, color=cut)) +
geom_smooth(mapping = aes(x = carat, y = price))
## Exercise 13: Coding for efficiency To reduce typing, put the aesthetics that are shared by all layers (global) in the ggplot() function. Put unique aesthetics in the geom() functions that are specific to that (local) layer only. Run this code to see what happens:
ggplot(data = diamonds, mapping = aes(x = carat, y = price)) +
geom_point(aes(color=cut)) +
geom_smooth()
Now use the short hand method to make the following code more efficient. Type your answer in the script:
ggplot(data = diamonds) +
geom_point(mapping = aes(x = carat, y = price, color=clarity)) +
geom_smooth(mapping = aes(x = carat, y = price))
ggplot(data = diamonds, mapping = aes(x=carat, y=price)) +
geom_point(aes(color=clarity)) +
geom_smooth()
It’s also possible to write it even more efficiently:
ggplot(diamonds, aes(carat, price)) + geom_point(aes(color=clarity)) + geom_smooth()
The previous chart has some issues with data points overlapping, and also too many spaghetti lines. We can use alpha, which is an aesthetic, and position to help reduce some of that overlap. Remember that position is one of the elements in the layered grammar of graphics.
alpha - use to make points more transparent so you can see points underneath position: takes values identity, dodge, fill, or jitter
ggplot(data =diamonds, mapping = aes(carat, price)) + geom_point(mapping = aes(color=clarity, alpha=1/5), position="jitter") + geom_smooth()
To make a scatter plot, we used the geom_point() function. You can use different geom functions to make other chart types. Here are just a few examples of the many geom functions:
geom_abline()geom_bar()geom_boxplot()geom_density()Each function can take certain parameters. To learn more about a function, you can type ?+name of function, for example, ?geom_bar
Use geom_bar() function to create a bar chart. Without a specified y variable, bar charts in ggplot calculate count, a new value. count is a statistical tranformation of your data that ggplot2 automatically does.
ggplot(data = diamonds) +
geom_bar(mapping = aes(x = cut))
Create a bar chart where x=clarity.
ggplot(data = diamonds) +
geom_bar(mapping = aes(x = clarity))
stat is another part of the grammar of graphics: essentially, a statistical tranformation of the data. Each layer has one.
geom functions each have their own default stat, or statistical transformation, measures, that transform the data. * geom_bar has a default stat called count * geom_histogram has bin …and so on
It’s possible to change these default settings, but normally, you don’t need to. Notice that we didn’t have to set stat to a value in order to create a bar chart.
To find out how to do this, type ?+package name in the console. Example ?geom_bar()